The thought of losing your photos is a nightmare and I was recently reminded of the importance of keeping an up-to-date working backup when, after some distinctly odd behaviour, my c:\ drive went down. Thanks to the excellent Live Mesh I had current copies of all my ongoing work files distributed around my network, and online too, in case the house/office burnt down. So that just left my photos…

If anything I’m even more paranoid about losing my photo collection, so I had full backups of all my original files and I use the free Google Picasa to organise them, so it was simple to download a copy of the latest version (Picasa 3.8 has just been released). However, faced with the need to restore, it hit me that a collection is far more than just the original photos and the management software – it’s also all the organisational information that makes the photos meaningful and accessible.

Now this wasn’t an entirely new thought, and nor was its significance for backup. One of the major reasons I had switched from Photoshop Elements to Picasa for organising my images was that I had grown suspicious of Elements’ all-your-eggs-in-one-basket approach, based on a central all-important catalogue that becomes larger and larger and so more and more vulnerable, as well as harder but more important to backup. It was a point that had hit home when my father’s Elements’ catalogue became corrupted and he was left scrabbling for his most recent backup disks.

By contrast Picasa takes a much more distributed approach. Rather than laboriously tagging every image as Elements expects, for example, you can name your folder with the date, people and event and then live search on these (plus embedded metadata, image caption, description etc). The folder-based approach isn’t as precise, but it’s far quicker to apply and surprisingly effective for retrieval. Moreover, because it’s more distributed, it’s also more robust – I knew that this core folder-based organisation would be recovered with the embedded metadata.

But what about the other organisation?  Suddenly I thought about all the work that you do in Picasa: creating albums, reordering, starring and rotating images, applying non-destructive edits and managing Picasa’s brilliant face tags. Clearly this information must be handled in a central database somewhere and presumably I had now lost it.

I needn’t have worried. Picasa does indeed have a central database, but it also cleverly distributes the information it needs to rebuild its database as hidden .picasa.ini files within each folder. Each ini file is a simple text file containing unsaved information on edits, stars and even the placement and identification of faces.

It meant that all I had to do was leave Picasa on overnight and, when I came in in the morning, Picasa had recreated its database and everything was up and running  as it had been before the crash. The only sign that this was a fresh set-up was that my name tags (some 10,000 spread across around 50 people) were now unnamed.  I named the first couple manually and then spotted something about logging in to web albums, which I did, and Picasa automatically named all the rest.

Belt and Braces Backup

I was more than happy with this, but it turns out that not quite all the information is stored in the picasa.ini files as, for some reason, these don’t store data for albums or image reordering.  To back this information up you must use Picasa’s Tools > Backup Pictures command.

With its prominent Burn button I had always assumed that this was intended solely for backing up to CD/DVD and, with 70GB of images, the thought had always been too awful to contemplate compared to the simplicity of drive-to-drive backup. In fact, digging in the Help file, I discovered that it’s perfectly possible to backup drive-to-drive, you just have to create a new Backup Set and the option becomes available. Although I’ll probably still stick to distributed folders rather than centrally-managed albums, I will certainly use this approach in future as all photos in the backup set are still simply stored in their folder structure (complete with ini files), but it makes it much easier to update only new and changed files.

Also while digging in the Help file, I came across some other important information. Another of Picasa’s great advantages is that it comes with 1GB of free web space for sharing web albums; what I hadn’t realised is that Google had radically cut its price on additional storage. With 80GB now costing only $20 a year (compared to $150 previously), and the ability to automatically synchronise original and changed files directly from Picasa, I’ve signed up and plan to add a Mesh-style online backup to my local backups.

Between its distributed database handling, disk-to-disk backup and bargain online storage, Picasa really does cover all bases and offers an excellent way to keep your images organised and safe.

It’s a maxim that’s been particularly on my mind lately, after I spent last month testing 21 external hard disks – almost all of which came with some sort of backup software – and then, this week, looked at two standalone backup applications as well. Right now, if there’s anyone who’s apprised of his backup options, it’s me.

But do you know which backup package I’ve chosen for my own use? None.

Keeping work files safe

Now, before you all bombard me with contemptuous comments, let me be clear. I’m not saying I don’t keep copies of important files. PC Pro is produced to a tight and immovable schedule, so none of us can afford to be blasé about the prospect of losing work. (Ask David Bayon, who once accidentally deleted the text of a twelve-page Labs section and had to pull some very late nights indeed to reproduce it in time for print.)

But I don’t need special backup software to make copies of my work files — because I save them into my Windows Live Mesh folder, from where they’re immediately beamed up to a remote server and synchronised with my laptop and home PC. In terms of data redundancy you really can’t do much better.

Ah, but, you may say: Live Mesh has comparatively limited storage, allowing you to synchronise a mere 5GB. What use is that in these days of terabyte drives?

What really needs backing up?

And that brings me to my main point. Yes, I would need a hundred Live Mesh accounts to back up my entire PC. But what exactly would I be backing up? The overwhelming majority of my hard disk space is taken up by videos, music, programs and system files — files I didn’t create myself and could easily replace.

(Yes, it’d be a pain rebuilding a music collection from scratch, but unless my MP3 player dies at the same time I can always just copy the files back from there.)

When you come down to it, if my computer were to wink out of existence tomorrow, the files I’d actually miss would add up to no more than a few gigabytes. And because the ones I’m actively working on are protected by Live Mesh (and are often wholly disposable once I’ve finished with them), the stuff that I need to back up myself changes by just a few megabytes a month.

Darien’s brute-force backup

So I have a batch file on my desktop which contains just one line:

This snappy command mirrors my user folder to a backup destination on another PC, recursing subdirectories but skipping NTFS junctions, system and hidden files, plus movies, music and VirtualBox disc images, files beginning with “_” (generally superfluous support files) and folders with names that start with “temp”.

The first time I ran it it took a good few minutes to do its work; but now I double-click on it whenever the mood takes me and, since it only copies changed files, it takes just a few seconds to update my backup. And if I ever need to rebuild my main machine, I can simply reinstall Windows, copy the “darien” folder back into its place and all my files (along with lots of useful application settings) are restored.

Of course, my system isn’t as versatile as a “real” backup package. It doesn’t run automatically, and it doesn’t encrypt or compress my personal data — not that it would be hard to add those capabilities if I could be bothered. And it doesn’t help me if I discover a month down the line that I’ve accidentally overwritten a vital file with gibberish.

But for all its limitations this single command has, so far, answered my backup needs admirably. I’m aware that it copies a lot of application support files that aren’t strictly necessary, but when you consider the wastage involved in a complete system backup I think it’s still a comparatively efficient approach.

Feel free to suggest refinements – or to tell me I’m an idiot, and point out what I should be doing instead. I’m interested to hear what you think.

And it won’t come with Internet Explorer 8, in a strangely “the sky is falling down” reaction to the antitrust spat with the EU. Microsoft hopes that this is enough, the EU doesn’t appear to be convinced.

But this could be a marvellous thing indeed. By forcing you to wipe the machine and start afresh, it is possible to ensure that all that accumulated crap is wiped from the machine. A clean OS is a happy OS.

But there’s more – wiping the machine means you need to have a working and tested backup/disaster recovery plan in place. If you don’t have one, go buy a USB hard disk and a copy of Norton Save & Restore 2. Do the backup that you have been meaning to do. Wipe the machine and install Windows 7, and then go back to Vista if there are problems.

But there’s even more – why not stuff some “cheap as chips” extra RAM into the computer and go for the 64-bit version? It really is time that we left the 32-bit world behind, and a machine wipe/reinstall is the ideal time to get into the 64-bit era

So despite all the squirming, handwringing and collective moaning and whining, I say “thank you” to Microsoft to force some end users to open their eyes, look about and then move forward. Pity they didn’t have the guts to do it to their American customers.

But I strongly advise you to go to http://ma.gnolia.com and watch the video. The background is that magnolia was a service which stored bookmarks. A lot of people relied upon it for their bookmark stores, and it was a “software as a service”. A cloud computing thing, if you like.

Unfortunately, it has gone bang, with irretrievable data loss.

Without question, this has been a learning experience for those enthusiasts who put a lot of effort into building the service, irrespective of their understanding (or lack of it) of risk assessment.

When you watch the video, you too will wince as I did about the choices they made. “Did you do any backup testing?” “Nope” made me squirm.

But think of the wider SaaS perspective. When you put a SaaS into your business process, do you really know who you are dealing with? Have you really looked at the SLA, the recovery tools and so forth?

Of course, it would be wrong to scale up a disaster at a community enthusiast site into the wider space of Azure or Google or Amazon services. But vendors will be selling SaaS services hosted on those platforms, and on all sizes of platforms all the way down to one PC with no working backup.

At the end of the day, it’s your business and you have to take responsibility for it. SaaS is not a major thing which cures all known ills.

But about three weeks ago, after a brutal personal slur on my character by PC Pro’s deputy editor David Fearon, I finally did it. “For [insert preferred deity here]’s sake Tim, we’ve got the software on the cover disc, just load it up and follow the wizard.”

The software is Carbonite, which has been our favoured online backup software for many a month now, and the reason the software’s on our cover disc is because we managed to negotiate six months’ worth of free storage for PC Pro readers rather than the 15 days you’ll get if you go straight to www.carbonite.com.

[I've just re-read that paragraph and it sounds like a plug - but it is a fantastic deal! Before people start flaming me, take note that I've never before blogged about pieces of software that we happen to have on our cover disc...]

Now the reason I title this blog “The joy of backup” is that I really am gaining some sort of strange feeling of happiness from seeing those green dots. I even know it works. As regular readers of this blog will know, I did a clean install of Windows 7 on my work laptop, and so I had to reinstall Carbonite – when I did, I was asked if I wanted to restore all my backed up data too. Nice and simple.

So I implore you, take advantage (the issue, which is the one that says “eBooks: The verdict” on large letters, is on sale for another two-and-a-half weeks, and the software is on both the CD edition and the DVD edition). You too will load up Windows Explorer and see reassuring green dots next to the folders and files that have been backed up. You too will get a strange feeling of wellbeing.

And before anyone points out that I should have a form of physical backup as well, fear not – I’ve got a lovely external hard disk tucked away with all the data safe and sound. 

There’s something in the air. In fact, look closely and you can see it. Up there, white and fluffy, and apparently full of exciting new technology. You see, all the best new technologies are moving to the cloud, and if you haven’t moved your data to the cloud yet… well, heaven forfend.

I’ve just deleted filed two press releases in the space of two minutes that sell themselves on the back of cloud computing. The first was from respected backup specialists EMC: “EMC has today announced its first major step into cloud computing as EMC Mozy and Lenovo will now offer unlimited online backup for SMBs” to quote the first line – head to Lenovo’s Protect My Data site for details.

The second is from F-Secure, which has just launched Wellbeing 2009, its new and trendier name for its suite of security products: Internet Security 2009, Anti-Virus 2009 and Home Server Security 2009.

Apparently, with this release “F-Secure is the only anti-virus vendor to have an ‘in the cloud’ real-time protection network deployed globally”.

Now I don’t dispute for a second the marvellousness of both these new sets of technologies, but the brutal reality is that the first has very little to back up its cloud claims. So far as I can see, it’s another online backup service (and don’t get me wrong, because Mozy is among the best).

This is in stark contrast to the F-Secure press release, where the cloud computing tag is completely justified. The central idea is that any PCs protected with F-Secure take a “fingerprint” of any files that are suspected of being malicious.

F-Secure’s ferocious fleet of servers then analyse the file and if it’s suspicious they’ll feed that information to all the other PCs protected by F-Secure’s technology (DeepGuard, to give it its marketing spin).

The problem is, cloud computing is just going to become another marketing term used without any great thought to describe any service delivered over the internet. And in the end, all that will happen is that people will get confused and ignore it.

Which is a shame, as real cloud computing is – as that F-Secure example shows – a very powerful thing.